A new automatic sleep stage classification model using swarm intelligence-based hybrid transfer learning architecture

Existing automatic sleep stage classification systems have mostly relied on hand-crafted features selected from polysomnographic records. To measure the quality of sleep, the automatic sleep stage classification system is very important. The sleep specialists examine the signals such as Electromyogr...

Full description

Saved in:
Bibliographic Details
Published in:Signal, image and video processing Vol. 18; no. 2; pp. 1131 - 1142
Main Authors: Raja, A. Ravi, Polasi, Phani Kumar
Format: Journal Article
Language:English
Published: London Springer London 01.03.2024
Springer Nature B.V
Subjects:
Accuracy
Classification
Computer Imaging
Computer Science
Datasets
Decomposition
Deep learning
Design
Effectiveness
Electrocardiography
Electroencephalography
Fourier transforms
Image Processing and Computer Vision
Machine learning
Multimedia Information Systems
Optimization
Optimization algorithms
Original Paper
Pattern Recognition and Graphics
Signal processing
Signal,Image and Speech Processing
Sleep disorders
Swarm intelligence
Vision
Sleep stage classification
Pre-processing using artifacts removal and band pass filter
Short-time fourier transform
Electroencephalograms signal
Optimal hybrid transfer learning approach
ISSN:1863-1703, 1863-1711
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Existing automatic sleep stage classification systems have mostly relied on hand-crafted features selected from polysomnographic records. To measure the quality of sleep, the automatic sleep stage classification system is very important. The sleep specialists examine the signals such as Electromyograms, Electroencephalograms (EEG), Electrocardiograms, and Electrooculograms, based on the visual inspection that is assigned every 30 s of the signal at the sleep stage. Hence, this research plans to implement an effective sleep stage classification model for detecting sleep disorder patients. It is performed with filtering approaches together with artifact removal techniques to get the pre-processed EEG signals. This pre-processed signal is used in the signal decomposition phase, where the short-time Fourier transform is involved in decomposing the pre-processed signals. Furthermore, these decomposed EEG signals are utilized in the optimal hybrid transfer learning approach for sleep stage classification using Mobilenet and Densenet techniques. The optimization takes place in the hybrid transfer learning approach with the development of a hybrid optimization strategy Hybrid Coyote Cat and Mouse Optimization Algorithm, to make efficient and accurate classification results. Experimental analysis reveals that the developed approach attains better effectiveness by analyzing various comparative techniques using different performance measures.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1863-1703
1863-1711
DOI:10.1007/s11760-023-02792-9